At present, the image recognition technology has been widely used in a variety of data processing systems, so that the data processing system can recognize a target object in an image and perform subsequent operations based on the recognition result.
For example, a device such as an automated teller machine may be installed with a camera. The camera transmits an acquired face image to a background server, and the background server performs identity identification based on the face image and performs operations related to withdrawing money based on the identity identification result.
It may be understood that, the quality of the image acquired by the camera may directly affect recognition efficiency and recognition accuracy of the background server. For example, whether the target object in the image acquired by the camera is complete, or whether an attitude of the target object meets requirements, may affect the image recognition effect of the background server. Therefore, it is desirable to provide a technical solution to ensure that the image acquired by the camera meets the image processing requirements.